Verify
Verify answers a different question from the read methods: not what does the chain say, but how far can this particular claim be believed.
Not available on this deployment. No capability row is commercially available for Verify here. This page describes what the surface is, because the vocabulary it uses appears in every other answer the API gives.
Five dimensions, never combined
There is no verified: true field anywhere in this API, and no
score. The useful answers are mixtures, and a single number cannot say which
mixture.
| Dimension | The question it answers |
|---|---|
| Trust class | What the answer rests on: the protocol itself, an outside attestation, or one source’s word. |
| Coverage | How much of what you asked about the answer actually covers. |
| Evidence basis | Where the evidence came from — including “XRPC’s own state”, which carries no protocol trust and never borrows any. |
| Read context | Whether several reads describe the same moment. |
| Native finality | What the chain itself says about settlement, in its own words. |
Why they stay apart
A source can be fresh and untrusted. An anchor can be strongly trusted and stale. A read can be complete about one thing and blind to another. Collapsing those into one flag is how monitoring stays green while the answers are wrong.
Verifiable is not verified. A claim that could be checked against the protocol is not a claim that has been. This API says which, every time, and refuses rather than guessing when it cannot.
Where you meet this vocabulary anyway
Every answer from every method carries a trust block using these terms, so the model matters even though the Verify product is not served here. See the trust model for what each value means and what you should do about it.